Test the hypothesis that some factor(s) interact(s) with menopause to predispose some but not other postmenopausal women to osteoporosis by comparing osteoporotic and control postmenopausal women before and after eliminating the confounding variable of estrogen deficiency by treatment.